Work with Data Wrangler

Data Wrangler is a no-code tool that simplifies data cleaning and preparation.

It offers an interactive user interface that allows you to view and analyze the data, displays column statistics and visualizations, and automatically generates Python code.

Open Data Wrangler

Open a Jupyter notebook.
Run code cell to create a pandas dataframe. For example, run cell with the following code:
import pandas as pd # Data data = { 'Name': ['John', 'Anna', 'Peter', 'Linda', 'Dina', 'Kate', 'Tom', 'Emily'], 'Age': [22, 78, 22, 30, 45, 30, 35, 40], 'Gender': ['Male', 'Female', 'Male', 'Female', 'Female', 'Female', 'Male', 'Female'], 'City': ['New York', 'Los Angeles', 'Chicago', 'Houston', 'Phoenix', 'Philadelphia', 'San Antonio', 'San Diego'], 'Occupation': ['Engineer', 'Doctor', 'Teacher', 'Nurse', 'Architect', 'Lawyer', 'Accountant', 'Scientist'] } # Create a DataFrame df = pd.DataFrame(data) # Display the DataFrame df
In the upper-right corner of the output cell, click Open Data Wrangler.
Data Wrangler will open in a new tab:

Use Data Wrangler transformations

Data Wrangler transformations in the Data Wrangler tab

Transformation	Description
Sort and filter
Filter	Filters rows in a selected column based on a specified condition and value Currently is not supported for string values
Clean and remove
Drop column	Removes a selected column from a table
Remove duplicates	Removes all rows that have duplicate values from a selected column
Drop missing values	Removes all rows with missing values from a selected column
Remove rows with NaN values	Removes rows that contain empty values from a table
Drop rows	Removes selected rows from a table
Find and replace
Find and replace	Replaces cells with a specified matching pattern from a selected column
Create additional column
Transform column with string	Transforms strings in a selected column. You can select one of the following transformations: Capitalize first character Convert text to lowercase Convert text to uppercase
One-hot encoding categorical variables	Splits categorical data from a selected column into a new column for each category
Other
Fill missing	Replaces cells with missing values with a new value in a selected column
Round numerical	Rounds numbers in a selected column to the specified number of decimal places: Round: rounds a number to the nearest integer. If the fraction of the number is 0.5 or higher, it rounds up. If it's less than 0.5, it rounds down. Floor: rounds a number down to the nearest integer Ceil: rounds a number up to the nearest integer
Split column	Splits a selected column into several columns based on a user defined delimiter
Change a type of column	Changes the data type of selected column
Normalize and scale
Min-Max scaling	Rescales a selected numerical column between a minimum and maximum value
Z-Score normalization	Transforms the data from a selected column into a distribution with a mean of 0 and a standard deviation of 1
Handling outliers or skewed distributions
Outlier detection with IQR	Detects outliers in a selected column using Interquartile Range
Reduce skewness	Reduces skewness by applying logarithmic or square root transformation to the data in a selected colum
Outlier detection with MAD	Detects outliers in a selected column using Median Absolute Deviation
Outlier detection with Euclidean distance	Detects outliers in a selected column using Euclidean Distance

Export Code to Notebook

You can create a new cell in your Jupyter Notebook with all the data transformation code you generated.

Click Export Code to Notebook.
You can view the history of changes applied to data before you export the code.
Your Jupyter notebook will open, and a new cell with generated code will be added to the notebook.

Export Data to File

You can save the transformed dataset as a new file.

Click Export Data to File.
Choose an extractor and configure additional settings.
Click Browse to choose the location for your file.
Click Export to File to save the data as a file.

Example: remove duplicate entries

One of the common data cleaning tasks is to remove duplicate entries to prevent biased results from your analysis.

You can use Data Wrangler to transform your data through the interface. Data Wrangler will automatically generate the Python code required for the removing of duplicates.

Open Data Wrangler.
Select Remove duplicates from the Transformations drop-down list.
Select the column from the Column drop-down list.
Check the generated code.
Click Apply.
You can see all the changes you applied to the data in the History pane.
Click Export Code to Notebook if you want to add a new code cell with generated code to your notebook or click Export Data to File to save transformed data as a file.

Last modified: 20 November 2024